This project is a basic project on hindi , gujarati text ocr using easyocr
The project is for detecting hindi,gujarati text in various images. It uses easyocr to get the text and bounding boxes coordinates.
The tech stack used for this project
- EASYOCR
- PYTHON3
- OPENCV
The input for the project is taken from https://www.kaggle.com/datasets/dataclusterlabs/hindi-text-image-dataset.
- To run the program on a single image , run
python3 task_1.py -i imagepath
- To run the program on a folder, run
python3 task_1.py -f folderpath
- Some output images are shown below
- For about 100 images using GPU from google colab , the time took was around 10 minutes .
- For a single image the time was 1 second.
- Some optimisations are yet to be done.
See the below links for more images and results. https://drive.google.com/drive/folders/14t1gW2RJVd9qD91BntOwvpf8diZsDn7Z?usp=sharing